signals-scout-general
Pass
Audited by Gen Agent Trust Hub on Nov 3, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted data from PostHog event streams, error logs, and session activity, which provides a surface for indirect prompt injection where malicious data could influence agent behavior.
- Ingestion points: Untrusted data is retrieved through tools such as
read-data-schema,query-trends,query-error-tracking-issues-list, andexecute-sql(SKILL.md, references/discovery.md). - Boundary markers: The instructions do not specify explicit delimiters or warnings for the agent to distinguish between its instructions and the data being analyzed.
- Capability inventory: The skill possesses write capabilities to the project's inbox via
scout-emit-reportandscout-edit-report, and can store persistent data in thescout-scratchpad-remember(SKILL.md). - Sanitization: The skill includes explicit security best practices in
references/new-traffic.mdinstructing the agent to use hashed identifiers (source_id) instead of raw event property strings in SQL queries to prevent injection, which mitigates technical injection risks but does not eliminate prompt-level influence from data content.
Audit Metadata